Evolutionary Ensemble of Agents

TL;DR

This paper presents EvE, a decentralized co-evolving system of coding agents that autonomously discovers robust algorithmic solutions, notably improving generalization in complex codebases.

Contribution

EvE introduces a novel evolutionary framework that co-evolves agent guidance and code solvers, enabling autonomous discovery of advanced algorithms and overcoming static performance limits.

Findings

EvE autonomously discovered a rescale-then-interpolate mechanism for ICON.

Stage-dependent agent adaptation is crucial for navigating complex code landscapes.

Organizing agents into a self-revising ensemble surpasses fixed or frozen agent variants.

Abstract

We introduce Evolutionary Ensemble (EvE), a decentralized framework that organizes existing, highly capable coding agents into a live, co-evolving system for algorithmic discovery. Rather than reinventing the wheel within the "LLMs as optimizers" paradigm, EvE fixes the base agent substrate and focuses entirely on evolving the cumulative guidance and skills that dictate agent behaviors. By maintaining two co-evolving populations, namely functional code solvers and agent guidance states, the system evaluates agents through a synchronous race, updating their empirical Elo ratings based on the marginal gains they contribute to the current solver state. When applied to a research bottleneck in In-Context Operator Networks (ICON), EvE autonomously discovered a robust rescale-then-interpolate mechanism that enables reliable example-count generalization. Crucially, controlled ablations reveal the absolute necessity of stage-dependent agent adaptation to navigate the shifting search landscapes of complex codebases. Compared to variants driven by a fixed initial agent or even a frozen "best-evolved" agent, EvE uniquely avoids phase mismatch, demonstrating that organizing agents into a self-revising ensemble is the fundamental driver for breaking through static performance ceilings.